A threonyl-tRNA synthetase-mediated translation initiation machinery

Abstract

A fundamental question in biology is how vertebrates evolved and differ from invertebrates, and little is known about differences in the regulation of translation in the two systems. Herein, we identify a threonyl-tRNA synthetase (TRS)-mediated translation initiation machinery that specifically interacts with eIF4E homologous protein, and forms machinery that is structurally analogous to the eIF4F-mediated translation initiation machinery via the recruitment of other translation initiation components. Biochemical and RNA immunoprecipitation analyses coupled to sequencing suggest that this machinery emerged as a gain-of-function event in the vertebrate lineage, and it positively regulates the translation of mRNAs required for vertebrate development. Collectively, our findings demonstrate that TRS evolved to regulate vertebrate translation initiation via its dual role as a scaffold for the assembly of initiation components and as a selector of target mRNAs. This work highlights the functional significance of aminoacyl-tRNA synthetases in the emergence and control of higher order organisms.

Fig. 1

Human TRS interacts specifically with eIF4E2 (4EHP). a Pull-down assay of co-expressed TRS-Strep and eIF4E-FLAG isoforms in 293T cells. TRS-Strep was pulled down with Strep-Tactin beads, and co-precipitation of eIF4E isoforms was analyzed by immunoblotting with anti-FLAG antibody. b Immunoassay of co-expressed 4EHP-FLAG with human ARSs in 293 T cells. 4EHP-FLAG was immunoprecipitated with anti-FLAG antibody, and co-precipitation of ARSs-Strep was determined by immunoblotting with anti-Strep antibody. c Endogenous TRS immunoprecipitation with anti-TRS antibody in WI-26 cell lysates and co-immunoprecipitation of endogenous 4EHP determined using anti-4EHP antibody. Rabbit IgG was used as a negative control. d Co-immunoprecipitation of the above proteins confirmed in the reverse order. Rabbit IgG was used as a negative control. e Interactions between the indicated protein pairs determined by reconstitution of Venus green fluorescent protein (GFP) in Chinese hamster ovary (CHO) cells. Nuclei were stained with 4′,6-diamidino-2-phenylindole (DAPI; blue). Expression of HA-4EHP and FLAG-TRS was confirmed by immunofluorescence staining with anti-HA (Alexa 647; purple) and anti-FLAG (Alexa 594; red) antibodies, respectively. Scale bar = 10 μM. f Domain structure of human TRS determined based on the crystal and solution structures of E. coli (PDB 1QF6) and human (PDB 1WWT) TRS. g Immunoassay of co-expressed 4EHP-FLAG with full-length GST-TRS or the indicated domains in 293 T cells. Cell lysates were immunoprecipitated with anti-FLAG antibody, and co-precipitated TRS domain(s) were determined by immunoblotting with anti-GST antibody. h ITC determination of the binding affinity and stoichiometry of the TRS UNE-T region and 4EHP. Raw data and the integration plot are displayed in the upper and lower panel, respectively. Data are representative of at least three experiments, each with similar results (a−e, g, h). EV empty vector, WCL whole cell lysate, VN venus N-domain, VC venus C-domain, UNE-T unique region extension at the N-terminus of TRS, TGS a domain named after TRS, GTPase and SpoT, ED editing domain, CD catalytic domain, ABD anticodon-binding domain

Fig. 2

The structure of the TRS UNE-T region bound to 4EHP implies that it is part of the translation initiation complex. a Overall structure of the TRS UNE-T-4EHP complex. The F_o − F_c electron density map of TRS UNE-T is displayed using surface representation of 4EHP calculated before the inclusion of TRS UNE-T, and contoured at 2.5 σ. The TRS UNE-T (magenta)-4EHP (pale cyan) complex is shown in cartoon representation. The disordered region in 4EHP (residues Pro69−Tyr78), which was not observed in the electron density map, is indicated by the dashed line. The asterisk indicates the cap-binding site in 4EHP. Hs, Homo sapiens. b Superimposition of the structures of the 4E-BP1 (cyan)-eIF4E (dark gray) and eIF4G (yellow)-eIF4E (light gray) complexes onto that of TRS (magenta) bound to 4EHP (pale cyan). The m⁷GTP cap analog bound to eIF4E is shown in stick representation. Dorsal and lateral surfaces are indicated. c Superimposition of the structures of the 4E-BP1 (yellow)-4EHP (dark gray), GIGYF1 (green)-4EHP (light gray), and GIGYF2 (blue)-4EHP (gray) complexes onto that of TRS (magenta) bound to 4EHP (pale cyan). d, e Interactions between TRS and 4EHP mediated by the canonical eIF4E/4EHP-binding motif. TRS residues (magenta) involved in the interaction with 4EHP are shown in stick representation (d). Surface region (pale cyan) of 4EHP residues responsible for interactions with TRS (d). eIF4E/4EHP-interacting proteins superimposed on the TRS-4EHP complex (e). Sequence alignment of the eIF4E/4EHP-binding consensus motif. Y and L/M of the consensus sequences (YX₄Lϕ) and Y of the additional N-terminal residues (YX) are highlighted by red and black dotted circles, respectively (e). f Features of TRS Tyr55 (left) and Met60 (right) interacting with 4EHP

Fig. 3

The TRS-4EHP interaction represents an evolutionary gain-of-function in vertebrates. a, b Structure-based sequence alignment of TRS and 4EHP in different species. Residues critical for the interaction between TRS and 4EHP are indicated by arrowheads. c−g Immunoassay of interactions between TRS and 4EHP pairs from the indicated species in 293T cell lysates (c, d, f, g) and Drosophila S2 cell lysates (e). Myc-TRS was immunoprecipitated with anti-Myc antibody, and co-precipitation of HA-4EHP was determined by immunoblotting with anti-HA antibody (c, f, g). HA-z4EHP was immunoprecipitated with anti-HA antibody, and co-precipitated zTRS WT or mutant (D50I, I55D) was determined by immunoblotting with anti-Myc antibody (d). V5-fTRS was immunoprecipitated with anti-V5 antibody, and co-immunoprecipitation of HA-f4EHP was detected by immunoblotting with anti-HA antibody (e). h In vitro pull-down assay of TRS (UNE-T)-His and GST-4EHP pairs from the indicated species. TRS-His was pulled down with Ni-NTA resin, and co-precipitated GST-4EHP was eluted from the resin and detected by Coomassie staining. Black and red arrowheads indicate eluted TRS (UNE-T)-His and GST-4EHP, respectively. Purified GST-fused 4EHP proteins from the indicated species are shown. Data are representative of at least three experiments, each with similar results (c−h). Hs Homo sapiens, Mm Mus musculus, Dr Danio rerio, Dm Drosophila melanogaster, Ce Caenorhabditis elegans, Sp Schizosaccharomyces pombe,. m mouse, h human, n nematode, y yeast, z zebrafish, f fly, EV empty vector, WCL whole cell lysate, M molecular size marker, S soluble fraction containing proteins expressed in E. coli, E proteins eluted from Ni-NTA resin

Fig. 4

TRS-mediated translation initiation control of mRNAs in vertebrates. a Workflow used to identify TRS-targeted mRNAs. Total mRNAs isolated from 293 T cells were precipitated with anti-TRS or anti-AlaRS antibodies and/or mock IgG, and RNA sequencing of precipitated transcripts was conducted. TRS-enriched RNAs were subtracted from those enriched in AlaRS and IgG groups, and the two RNA pools were compared to identify common transcripts. Among the transcripts commonly detected in both TRS-enriched pools, 2,928 ≥1.75-fold were selected, and subsequently subjected to functional annotation. b, c Functional annotation of TRS-targeted mRNAs. Enriched GO terms in the Biological Process category were analyzed using the Database for Annotation, Visualization and Integrated Discovery (DAVID). d RNA immunoprecipitation of endogenous TRS in 293T cells followed by reverse transcription-PCR using primers within the 5′-UTR of VEGF mRNA. IN, input; nt, nucleotide. e Potential tRNA^Thr anticodon triplet (UGU)-containing stem-loop structures were observed at positions -167 and -749 upstream from the VEGF mRNA initiation codon. The two RNA sequences spanning positions -1 to -540 (5′-UTR-167) and -541 to -1,038 (5′-UTR-749) were fused upstream of the firefly luciferase gene (Fluc) and co-expressed with the Renilla luciferase gene (Rluc) in TRS- or AlaRS-expressing 293T cells. Data are presented as the ratio of firefly to Renilla luciferase activity (Fluc/Rluc). f Immunoblot analysis of TRS-Strep and AlaRS-Strep in TRS- or AlaRS-expressing 293 T cells using Strep antibody. * indicates a nonspecific band. g Translation of the pseudo-anticodon-containing reporter gene in siTRS- or siAlaRS-transfected 293 T cells. siCont, non-targeting control siRNA. h The effect of knock-down of TRS or AlaRS with its specific siRNAs was determined by immunoblotting with each antibody. *p < 0.05; **p < 0.01; NS not significant vs. control group. Values are means ± SD of three independent experiments (e−h). EV empty vector

Fig. 5

The 4EHP and TRS interaction is critical for translation initiation of mRNAs required for vascular development. a Effects of TRS and/or 4EHP on VEGF secretion in siTRS- and/or si4EHP-transfected cell lines. VEGF protein levels in the culture supernatant were determined by enzyme-linked immunosorbent assay (ELISA). b Effects of TRS and/or 4EHP on ANG secretion in siTRS- and/or si4EHP-transfected cell lines. ANG protein levels in the culture supernatant were determined by ELISA. c, d Effects of TRS and/or 4EHP on endothelial cell tube formation. Culture medium from WI-26 cells transiently transfected with siRNAs against TRS and/or 4EHP (or siCont) was used to treat HUVECs, which were subsequently plated on growth factor-reduced Matrigel to form capillary tubes (c). Total tube lengths in Supplementary Fig. 6a were measured using ImageJ (c). Migrated HUVECs were counted from three randomly selected fields (c). Culture medium from WI-26 cells transiently transfected with Myc-TRS WT or its 4EHP-binding-defective M60K mutant was used to treat HUVECs, which were subsequently plated on growth factor-reduced Matrigel to form capillary tubes (d). Total tube lengths in Supplementary Fig. 6f were measured using ImageJ (d). Migrated HUVECs were counted from three randomly selected fields (d). VEGF (10 ng mL^-1) served as a positive control. siCont, non-targeting control siRNA. *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001 vs. control group. Values are means ± SD of three independent experiments (a−d). e−h TRS- and 4EHP-mediated translation initiation is critical for vascular development. TRS was suppressed using a morpholino (trs i6e7 MO) together with control (Cont) RNA, or reconstituted with WT or I55D mutant trs RNA at 52 h post-fertilization (e). 4EHP was suppressed using a morpholino (4ehp MO) together with Cont RNA, or reconstituted with WT 4ehp RNA (g). Quantitation and statistical analysis of the lengths and branching points of central arteries in the hindbrain of Tg (kdrl:EGFP) zebrafish embryos (f, h). Scale bar = 100 μM. The number of analyzed zebrafish embryos is shown in parentheses. ***p < 0.001; NS not significant vs. control group. Values are means ± SD. SF serum-free medium

Fig. 6

TRS functions similarly to eIF4G and acts as an eIF4F analog. a Pull-down assay of co-expressed TRS-Strep with eIF4A- or eIF4G-FLAG in 293 T cells. TRS-Strep was pulled down with Strep-Tactin beads, and co-precipitation of eIF4A or 4 G was determined by immunoblotting with anti-FLAG antibody. EV, empty vector. * indicates a nonspecific band. b Immunoassay of the co-expression of different combinations of plasmids in 293T cells. Myc-TRS was immunoprecipitated with anti-Myc antibody, and co-precipitation of other proteins was determined using tag-specific antibodies. c Immunoassay of co-expressed eIF4A-FLAG with GST-fused full-length TRS or its various domains in 293T cells. eIF4A-FLAG was immunoprecipitated with anti-FLAG antibody, and co-precipitated TRS proteins were determined by immunoblotting with anti-GST antibody. d Pull-down assay of endogenous translation initiation factors with m⁷GTP-Sepharose beads in 293 T cells transfected with siRNAs against TRS, 4EHP, or a non-targeting control (siCont). Cap-bound proteins were eluted from beads and immunoblotted with the indicated antibodies. Sepharose beads were used as a negative control. e Pull-down assay of endogenous translation initiation factors with m⁷GTP-Sepharose beads in 293T cells transfected with siRNAs against eIF4G, eIF4E, or siCont, and their suppression effects on cap-binding of other components, were determined as in (d). The data are representative of at least three experiments, each with similar results

Fig. 7

Discovery of a TRS-mediated translation initiation machinery. a Pull-down assay of GST-TRS with FLAG-tagged proteins. Purified GST-TRS from E. coli was incubated with 293T cell lysates with FLAG-tagged 4EHP, PABP, eIF4A, or eIF4G. GST-TRS was pulled down with glutathione-Sepharose beads, and co-precipitation of each FLAG-tagged protein was determined by immunoblotting with anti-FLAG antibody. b Immunoassay of co-expressed PABP-FLAG with WT or M60K mutant TRS-Strep in 293T cells. PABP-FLAG was immunoprecipitated with anti-FLAG antibody, and co-precipitated TRS proteins were determined by immunoblotting with anti-Strep antibody. c Immunoassay of co-expressed PABP-FLAG with GST-fused full-length TRS or its various domains in 293T cells. PABP-FLAG was immunoprecipitated with anti-FLAG antibody, and co-precipitated TRS proteins were determined by immunoblotting with anti-GST antibody. d Endogenous TRS was immunoprecipitated from WI-26 cells with a mouse anti-TRS antibody, and co-immunoprecipitates were examined with the indicated antibodies. Mouse IgG was used as a negative control. e Pull-down assay of GST-TRS with FLAG-tagged eIF3 subunits. Purified GST-TRS from E. coli was incubated with 293T cell lysates containing each of the FLAG-tagged eIF3 subunits. GST-TRS was pulled down with glutathione-Sepharose beads, and co-precipitation of each FLAG-tagged protein was determined by immunoblotting. f Endogenous TRS was immunoprecipitated from WI-26 cells with a mouse anti-TRS antibody, and co-immunoprecipitates were examined with the indicated antibodies. Mouse IgG was used as a negative control. Data are representative of at least three experiments, each with similar results (a−f). g Structural model showing the interaction of 4EHP with the UNE-T region of the TRS dimer. Each domain of TRS is labeled as in Fig. 1f. A monomer of TRS is shown in surface representation, and the UNE-T regions interacting with both 4EHP molecules are shown in cartoon representation. The model was built based on the crystal structure (PDB 1QF6) of E. coli TRS. h Schematic model of the TRS-mediated translation initiation machinery. See the main text for a full description. Although TRS is predicted to function as a homodimer, the model only shows the monomer for simplicity. eIF3 is depicted by a dotted circle because its binding region in TRS has not yet been determined